This is a prospective single arm phase II trial of 60 stage IB NSCLC non-diabetic patients who will receive SBRT for definitive management. Subject selection will be limited to cases with available pathology for tumor mutation analysis. Study subjects will receive induction metformin alone (3 weeks) and concurrent metformin with SBRT (1 week). Response to therapy is the primary study outcome and will be assessed after induction and at study completion using [18F]-FDG PET/CT imaging (Fig. 4). The imaging response to induction metformin and combined SBRT plus metformin will be compared using RECIST version 1.1 (Eisenhauer, Therasse et al. 2009) versus PERCIST 1.0 (Wahl, Jacene et al. 2009) response criteria. Multiple expert observers will be utilized to perform repeat measurements for estimation of response variance. Pre-enrollment biopsy specimen tumor mutation analysis will evaluate candidate genes including TP53, Kras, STK11 using semiconductor sequencing (Rothberg, Hinz et al. 2011). Tumor mutation status will be characterized from pre- enrollment biopsy specimens for the candidate genes (TP53, Kras, LKB1) using semiconductor sequencing (Rothberg, Hinz et al. 2011). Statistical models will be developed to predict tumor response (PERCIST v1.0 & RECIST v1.1) to metformin induction and to combined metformin plus SBRT using pre-treatment characteristics - tumor glucose uptake and mutation status. We will estimate the area under the receiver operator characteristic curve (AUROC) for models comprising tumor genotype (tumor mutation) and phenotype (pretreatment SUV) each independently and combined. We will compare models for predictive power and simplicity (Akaike 1974).